LLM cost calculators
15 free calculators running on live pricing for 60 models across 17 providers. No signup, no account, and nothing you enter leaves your browser. Rates verified 2026-08-16.
Grouped by the question you are actually asking. If you are not sure where to start, the API cost calculator answers "what does one call cost?" and everything else builds on that.
Start here: what will this cost?
If you know roughly what your requests look like, these turn that into a number. Most people need only the first one.
LLM API Cost Calculator→
Estimate the cost of a single API call across any model.
Cost Per Request Calculator→
Price a single API call and scale it to per-1,000 and per-month totals.
Token Counter→
Estimate how many tokens a piece of text will use.
Tokens Per Dollar Calculator→
Flip the question: how many tokens does your budget actually buy?
Price a specific workload
Each of these bakes in the shape of a particular application, so you enter the parameters that matter for it rather than raw token counts.
Chatbot Monthly Cost Estimator→
Estimate what a production chatbot will cost per month.
RAG Pipeline Cost Estimator→
Estimate monthly cost of a retrieval-augmented generation pipeline.
AI Agent Cost Calculator→
Estimate the cost of multi-step, tool-calling agent runs.
Embeddings Cost Calculator→
Estimate the cost of embedding a document collection.
Context Window Cost Calculator→
See what it costs to fill a model's full context window.
Find savings
The levers that reduce a bill without changing what your product does. Run these before considering a model downgrade.
Prompt Caching Savings Calculator→
See how much a cached system prompt saves at your cache hit rate.
Batch API Savings Calculator→
Work out what moving a latency-tolerant workload to the batch API saves.
Model Switching Savings Calculator→
Compare two models on the same workload and see the annual difference.
Beyond text generation
Training, images, and running open-weight models on your own hardware.
Fine-Tuning Cost Calculator→
Estimate training and inference costs for a fine-tuned model.
AI Image Generation Cost Calculator→
Compare per-image costs across leading image models.
GPU Rental Cost Calculator→
Estimate the cost of self-hosting an open-weight model on rented GPUs.
What a calculator can and cannot tell you
Every tool here is arithmetic over published list prices. That makes them exact about the thing they measure and silent about everything else: none of them says whether a model is good enough for your task, how many retries you will need, or how much longer one model's answers run than another's. Those are the variables that most often make a real bill diverge from a projection, and the only way to pin them down is to run a sample of real traffic through the candidates.
Use the numbers as the budget for that evaluation rather than as its conclusion. A forty-dollar-a-month difference does not justify a week of migration work; a forty-thousand -a-year difference justifies quite a lot of it. For the rates behind every calculation, see the every model compared on price, and for the reasoning behind per-token billing itself, the LLM API pricing.
Questions about the calculators
Are these LLM cost calculators free?
All fifteen, with no signup and no account. They run entirely in your browser, so nothing you type is sent to a server or stored anywhere.
Which calculator should I use first?
The API cost calculator if you know your token counts, or the token counter if you need to estimate them from real text first. Everything else answers a narrower question once you have that baseline.
How accurate are the results?
The arithmetic is exact and the rates are current as of 2026-08-16. What makes real bills differ from an estimate is almost always the inputs: retries, prompt growth over time, and outputs longer than planned. Estimate on your observed p50 and p90 token counts rather than your intended ones.